INDEX

Explanations

bottom and right spacing

This neuron activates on numeric literals (especially CSS dimension values like “1400”, “362”, “83.5”, etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 alrededores

-1.28

 availa

-1.21

 MULTIPLE

-1.15

you

-1.10

 môžete

-1.09

 musíte

-1.09

 různých

-1.07

㩒

-1.07

pagest

-1.07

 mijne

-1.05

POSITIVE LOGITS

</tr>

1.20

 haja

1.06

 before

1.04

));

1.04

ers

1.04

</table>

1.03

 Conclusions

1.03

 борьбе

1.01

ίας

1.00

Activations Density 0.002%